KR100893718B1 - Smart Antenna with Adaptive Convergence Parameter - Google Patents

Abstract

The smart antenna (FIG. 2), i.e. blind adaptive antenna array, suppresses multiple access interference and includes CDMA radio including third generation (3G) CDMA2000 and wideband (W) -CDMA. A method and system for improving the performance of a communication system. Convergence parameters are employed in the smart antenna processor 207. In general, a constant convergence parameter value is empirically determined and used after reviewing the convergence rate and other capabilities such as steady state MSE or bit error rate. The convergence parameter yields crude performance when the channel environment changes (the channel environment actually changes because the mobile user generally walks around). In the present invention, the convergence parameter values are adaptively changed and employed in the smart antenna processor 207. Two example methods of updating the convergence parameter are described. By employing such adaptive convergence parameter values, the convergence rate can be increased and the steady state MSE can be reduced.

Smart Antenna, Smart Antenna Processor, CDMA, Convergence Parameters, Adaptive Convergence Parameters.

Description

Smart Antenna with Adaptive Convergence Parameter

The present invention relates to wireless communication. More specifically, the present invention relates to a new and efficient smart antenna for code division multiple access wireless communication systems. If the smart antenna of the present invention is employed in a base station and a desired mobile user moves around tall buildings, the smart antenna will work more efficiently than other existing smart antennas.

In third generation (3G) wireless communication systems, specifically those specified by the Wideband (W) -CDMA or CDMA2000 standard, the base station has the option to employ smart antenna technology. Smart antennas can suppress interferece signals of different directions of arrival angles (DOAs) from desired users by using spatial diversity. Smart antenna technologies have attracted a lot of attention recently because they support more users with higher quality of service and higher data rates such as 1.92 Mbps. Recently, efficient smart antenna schemes have emerged. One of them is invented by the inventors and is pending in US Ser. No. 60 / 164,552. It would be reasonable to include many practical environments not considered in the existing literature or inventions.

For example, if a mobile user is roaming downtown in a dense city of tall buildings, the direction of arrival angles (DOA) from the multipath signals of the desired mobile user may be locally scattered around the mobile user. Local scatters may cause a sudden change in the base station receiver. This phenomenon is called "edge effects". One of the roles of the smart antenna at the base station is to track the DOAs of the dominant multipath signal. Especially in the case where the above mentioned edge effects occur frequently, the convergence speed and DOA tracking capability of smart antenna weighting coefficients are very important issues in the design of smart antenna. (critical issues). Most of the existing literature or inventions on smart antennas do not include these edge effects, and have traditionally employed a constant convergence parameter. There is still a need for smart antennas with fast convergence speed as well as small mean square error (MSE) under edge effect environment. Some existing literature considers time-varying adaptive convergence parameters, such as, for example, normalized least mean squares (NLMS). U. S. Patent No. 5,999, 800 (Choi) also employs a time-varying LaGrange multiplier. However, these existing documents or inventions employ optimization standards other than the present invention or adaptation methods different from the present invention.

Smart antenna, a blind adaptive antenna array, suppresses multiple access interference and improves the performance of wireless communication systems including CDMA such as 3G (3G) CDMA2000 and W-CDMA And system. Some parameters are employed in the smart antenna processor. In general, a constant convergence parameter value is empirically determined and used after reviewing the convergence rate and other capabilities such as steady state MSE or bit error rate. As the convergence parameter value increases, the convergence rate also increases, but unfortunately the MSE increases and vice versa. Smart antennas with constant convergence parameters yield crude performance when the channel environment changes (the channel environment actually changes as mobile users generally move around). In the present invention, the convergence parameter values are adaptively changed and employed in the smart antenna processor. Two example methods of updating the convergence parameter are described. By employing such adaptive convergence parameter values, the convergence rate can be increased and the steady state MSE can be reduced. Simulation test results show improved performance for a CDMA system in which a smart antenna using the adaptive convergence parameter schemes of the present invention compared to existing schemes operates under time-varying fading channel and edge effects. It was confirmed. In addition, smart antennas according to the present invention have smaller computational loads than competing inventions such as Choi.

In accordance with the present invention, a method is provided for receiving a signal for use in combination with wireless communication. The present invention includes receiving a signal at multiple antennas. The received signal is processed using an adaptive convergence parameter. According to one embodiment of the invention, the antennas are a multiple antenna array. According to another embodiment of the invention, the antennas are multiple antennas.

According to one embodiment of the invention, the received signal is processed according to the following.

According to an alternative embodiment, the signal is processed according to the following.

However, the present invention is not limited to these two specific algorithms, which are merely exemplary.

The processing step includes estimating the DOA for the antennas, which is separated from other signal data. It also includes determining a better weighting coefficient.

According to one alternative, the DOA is used for forward link transmission. According to another alternative, DOA is used for reverse link transmission. Other alternatives include antennas at a base station, or mobile station.

In accordance with the present invention, there is further provided a signal receiving system for use in combination with wireless communication. The system includes at least one signal processor responsive to a signal received at one or more antennas with adaptive convergence parameters. Optionally, the system includes a transmitter coupled to the signal processor. Optionally, the system includes a receiver coupled to the signal processor. Optionally, the antennas comprise multiple antennas. Optionally, said signal processor comprises a filter, said filter having an adaptive convergence parameter.

Optionally, the signal processor includes a measurement of a DOA for the antennas, wherein the measured DOA is separated from other signal data. In the alternative, the signal processor further includes a determination of a better weighting factor. Alternatively, the measured DOA is obtained from reverse link transmission.

According to one embodiment, the system comprises a base station and the antennas are in the base station. According to yet another embodiment, the system comprises a mobile station and the antennas are in the mobile station.

Various objects, features and advantages of the invention will be readily apparent from the following drawings and detailed description.

The features, objects, and advantages of the present invention will become more apparent from the detailed description set forth below in connection with the drawings identified in correspondence with reference numerals.

1 is an exemplary model of a complex PN complex spread transmitter for a CDMA system constructed in accordance with one embodiment of the present invention.

2 is an overall block diagram of a base station receiver constructed in accordance with one embodiment of the present invention.

3 shows a flow chart of a smart antenna processor with an optimal convergence parameter algorithm constructed in accordance with one embodiment of the present invention.

4 shows a flow chart of a smart antenna processor with a heuristic convergence parameter algorithm constructed in accordance with one embodiment of the present invention.

Fig. 5 shows some simulation DOA estimation results for single realization and iteration, in which smart antennas constructed in accordance with one embodiment of the invention have a faster convergence rate and a smaller mean square. error).

Figure 6 shows two curves separated from Figure 5 for two adaptive convergence algorithms of the present invention.

Fig. 7 shows some simulated bit error rate (BER) results for the number of edge effect occurrences in a frame of 20 ms, which shows that a smart antenna constructed in accordance with one embodiment of the invention shows slightly better performance under edge effects. Indicates.

Although the present invention can be equally applied to a forward link beam-forming process as in a CDMA wireless communication system, the present invention contemplates a reverse link from a mobile station to a base station. In addition, although the transmitter and receiver of this embodiment are similar to the CDMA2000 system for demonstration purposes, the present invention relates to a W-CDMA system when the weight vector of the smart antenna is adaptively generated with a convergence parameter. The same may also apply to other CDMA systems.

은 1 또는 알려진 패턴의 ±1의 시퀀스이고, 트래픽 채널(traffic channel)에 있는 Q-채널 입력 데이터 스트림(105)

은 ±1의 랜덤 시퀀스이다. 여기에서 k는 코드 심볼 인덱스를 나타내고 n은 사용자 인덱스를 나타낸다. 파일럿 진폭(103) A₀는 1로 설정되고 왈시 코드(107)

은 ±1로 설정된다. i는 칩 인덱스(chip index)를 나타낸다. 각각의 코드 심볼은 G 칩들로 스프레드되고, 여기에서 G는 스프레딩 팩터(spreading factor, SF)로 불린다. I 및 Q 데이터는 복소 PN 스프레드 신호(108)

이다. PN 스프레드 신호는 다음과 같이 쓸 수 있다. 1 is an exemplary transmitter spread by a complex PN sequence for a CDMA system such as CDMA2000 or W-CDMA. I-channel input data stream 101 in pilot channel

Is a sequence of 1 or ± 1 of a known pattern, and the Q-channel input data stream 105 in the traffic channel

Is a random sequence of ± 1. Where k represents a code symbol index and n represents a user index. Pilot amplitude (103) A ₀ is set to 1 and Walsh code (107)

Is set to ± 1. i represents a chip index. Each code symbol is spread over G chips, where G is called the spreading factor (SF). I and Q data are complex PN spread signals 108

to be. The PN spread signal can be written as

(1)

(One)

와 같이 표현된다. n번째 사용자로부터의 전송된 신호(111) s_n(t)는 다음과 같이 쓸 수 있다.j is the positive square root of (-1). The same lowpass signal after pulse shaping filter 109 H (f)

It is expressed as The transmitted signal 111 s _n (t) from the nth user can be written as follows.

(2)

(2)

t is the time variable, Re {z} is the real part of complex z, P is the transmitted power, f _c is the carrier frequency, e is the exponential operator, and T _c is the chip time interval.

2 shows an example receiver block diagram with a smart antenna for CDMA reverse link. The present invention improves the performance of the smart antenna of FIG. The number of antenna array elements 201 is M. The array element spacing d is set to λ / 2. Where λ is the same wavelength as the speed of light divided by the carrier frequency. Since the maximum distance between the elements is (M-1) λ / 2 or less (31.6 cm when the carrier frequency is 1.9 GHz and M = 5), the signals arrive at each element substantially equally. The relative phase difference according to the array element spacing between the reference and mth element outputs is

θ는 희망 신호의 DOA이다. 상기 제1 요소는 레퍼런스 요소로 설정된다. 상기 안테나 배열 응답 벡터 a(θ)는 다음과 같이 쓰여질 수 있다.

θ is the DOA of the desired signal. The first element is set as a reference element. The antenna array response vector a (θ) may be written as follows.

(3)

(3)

T is a transpose operator.

The received signals (m = 1, ..., M) at the mth element can be written as

(4)

(4)

τ _{l, n} , α _{l, n} (t), φ _{l, n} (t) and θ _{l, n} (t) are the multipath delay, amplitude, phase and DOA of the lth path from user n, respectively. And n _{m, l} (t) is additive white Gaussian noise (AWGN) 203. The output of each element is frequency down converted. Baseband filter outputs 205 are sampled every chip interval and denoted by x _{l, m} (i). The smart antenna processor 207 takes pre-PN processing chip vectors x _l (i) and post-PN processing chip vectors y _l (i). Pilot-aided channel estimates 211 are obtained by taking the average of post PN despread 209 samples over an interval of N _p chips. Spatial and temporal RAKE combining 213 is performed on m = 1, ..., M antenna elements and l = 1, ..., L fingers. If a convolutional encoder is employed at the transmitter, the soft decision variable u (k) 221 of Figure 2 is a Viterbi convolutional decoder (Viterbi) with the soft decision 217 or hard decision 219 value for the kth code symbol decoding. convolutional decoder 215.

Preferably, the present invention employs a smart antenna based on maximum output power criteria as used in Choi. Other standards may be used as long as the smart antenna weight vector is updated with an adaptive algorithm. In Choi or other existing algorithms a constant convergence parameter μ is used.

The cost function based on the maximum output power standard can be written as

(5)

(5)

로 업데이트되고 여기에서

는 w(k)에 대해 방정식(5)에 주어진 코스트 함수 J(w(k))에 대한 그레디언트 벡터(gradient vector)이고, μ는 수렴 파라미터이다. 상기 업데이트된 웨이트 벡터는 다음과 같이 쓸 수 있다. w ^H (k) is the Hermition of the M by 1 weight vector w (k), ie conjugate and transpose, and y (k) is the post PN deprecated M by 1 vector (post PN despread). observed M by 1 vector), the mth element y _m (k) is the sum of the post PN spread samples y _m (i) in FIG. 2 on the G chip intervals, and R _yy (k) is y (k) ) Is the M by M auto-correlation matrix, and γ (k) is the LaGrange multiplier for the constraint w ^H (k) w (k) = 1. Finger index l is omitted later for brevity. K denotes the update time index and is called an iteration or snapshot index. In the preferred embodiment, although the update rate may be faster than the code symbol rate, the update rate is set to the code symbol rate. In each iteration, or snapshot, the weight vector

Updated to here

Is the gradient vector (vector gradient) for a given cost function J (w (k)) in the equation (5) for w (k), μ is a convergence parameter. The updated weight vector can be written as follows.

(6)

(6)

는 적응성 필터 출력이고 R_yy(k)는 y(k)y ^H(k)로 근사화된다.(만일 라그랑지 곱수 γ(k)가 상수값 γ를 사용하는 대신에 제약 조건(constraint) w ^H(k)w(k)=1 하에서 매 스냅샷마다 업데이트된다면, γ(k)는 반복이 계속됨에 따라 R_yy의 최대 고유값으로 수렴된다.)

Is an adaptive filter output and R _yy (k) is approximated to y (k) y ^H (k) (if the Lagrange multiplier γ (k) uses a constant value γ, the constraint w ^H ( k) If w (k) = 1 is updated every snapshot, then γ (k) converges to the maximum eigenvalue of R _{yy as} the iteration continues.)

Conventionally known adaptive filters empirically determine the convergence parameter μ in equation (6) as a constant. Determining the convergence parameter μ is a difficult process if the channel environment continues to change. The convergence parameter value μ affects the convergence speed of the adaptive algorithm. If μ is small, the convergence rate is low but the excess of MSE is small and vice versa. Preferably during every snapshot instead of always employing a constant convergence parameter, the present invention adaptively updates the convergence parameter μ (k). The present invention fixes Lagrangian multiceptor γ (k) to a constant value equal to the maximum eigenvalue of R _yy , ie γ (k) = M. This is because the adaptive γ (k) does not change the convergence speed. Thus, the cost function of equation (5) changes as follows for the present invention.

(8)

(8)

And the new weight vector w (k + 1) in equation (6) is

(9)

(9)

를 구하고 그것을 다음의 업데이트 방정식에 대입한다.:

또한, μ(k)에 대해 방정식(8)에서 코스트 함수의 미분을 취하고 그것을 0으로 함으로써, 코스트 함수를 최소화하기 위한 최적 적응성 수렴 파라미터 μ(k)가 다음과 같이 구해진다.Gradient

And replace it with the following update equation:

Further, by taking the derivative of the cost function in equation (8) with respect to μ (k) and making it zero, the optimal adaptive convergence parameter μ (k) for minimizing the cost function is obtained as follows.

(10)

10

는 옵저브드 post PN 디스프레드 백터(observed post PN despread vector) y(k)의 내적(inner product)이다.Where | z (k) | is the magnitude of the complex array output z (k)

Is the inner product of the observed post PN despread vector y (k).

The heuristic adaptive convergence parameter μ (k) is obtained as follows.

(11)

(11)

는 작은 전력을 갖고 상기 적응성 수렴 파라미터 μ(k)는 클 것이고 수렴 단계는 반복 k에서 크고 검색 처리는 속도가 빨라질 수 있다. 웨이트 벡터 w(k)가 채널 배열 응답 벡터 a(θ(k))와 매칭하는 경우, 상기 배열 출력

는 M²과 같은 최대 출력 전력을 갖고 적응성 수렴 파라미터 μ(k)는 작아지고 초과 MSE는 작아질 것이다. The heuristic adaptive convergence parameter of equation (11) makes sense. The array output if the weight vector w (k) does not match the channel array response vector a (θ (k)) of equation (3)

Has small power and the adaptive convergence parameter μ (k) is large and the convergence step is large at the iteration k and the search process can be speeded up. The array output if the weight vector w (k) matches the channel array response vector a (θ (k))

Has a maximum output power equal to M ² and the adaptive convergence parameter μ (k) will be small and the excess MSE will be small.

The convergence of the smart antenna of the present invention can be confirmed by examining Equation (9). The update weight vector can be rewritten as:

(12)

(12)

(13)

(13)

(14)

(14)

(15)

(15)

라면 반복이 계속됨에 따라 대각선 값들은 감소한다. 그래서, γ는 λ_max=M으로 설정된다.Where Q is a unitary matrix that satisfies R _yy = QΛQ ^H , Λ is a diagonal matrix with the i-th diagonal element equal to the i-th largest eigenvalue of the matrix R _yy , and w (0) Is the initial weight vector set to (1,1, ..., 1) ^T. The (k + 1) squared parenthesis matrix of equation (15) is a diagonal matrix

If it is, the diagonal values decrease as the iteration continues. Thus, γ is set to λ _max = M.

로 나타내어 지고 여기에서 b₁ 및 a(θ₁)은 각각 ±1의 데이터 비트 및 사용자1로부터의 DOA θ₁의 배열 응답 벡터이고, n(k)는 방해 플러스 열 잡음 벡터(interference plus thermal noise vector)이다. 희망되는 적응성 필터 출력 또는 레퍼런스 신호는 위너 필터에 대해

으로 설정될 수 있다. 이상적으로는, 크로스 코릴레이션 벡터(cross correlation vector) p는 다음과 같이 쓸 수 있다.For comparison purposes, the Wiener filer is again reviewed for smart antenna applications. M by 1 PN-spread output vector

Where b ₁ and a (θ ₁ ) are the data bits of ± 1 and array response vector of DOA θ ₁ from User 1, respectively, and n (k) is the interference plus thermal noise vector )to be. The desired adaptive filter output or reference signal can be

Can be set. Ideally, the cross correlation vector p can be written as

(16)

(16)

(17)

(17)

(18)

(18)

Then, an ideal Wiener solution can be obtained as follows.

(19)

(19)

(20)

20

(21)

(21)

(22)

(22)

(23)

(23)

(24)

(24)

(25)

(25)

(26)

(26)

이 사용되었다. 방정식(26)에 있는 배열 응답 벡터 a(θ₁)은 페이딩 위상 왜곡을 보상하고 DOA 콤포넌트들(DOA components)을 각각의 요소에만 유지시키기 위해 파일럿 채널 추정의 켤레(conjugate)를 곱함으로써 얻어질 수 있다. 위너 필터에 대한 시뮬레이션 테스트 결과들은, 최적 적응성 수렴 파라미터에 대한 방정식(9) 및 (10)의 쌍과 발견적 적응성 수렴 파라미터 알고리즘에 대한 방정식(9) 및 (11)의 쌍을 사용함으로써 본 발명으로 얻어진 결과들과 비교된다. In equation (22)

This was used. The array response vector a (θ ₁ ) in equation (26) can be obtained by multiplying the conjugate of the pilot channel estimate to compensate for fading phase distortion and keep the DOA components only on each element. have. Simulation test results for the Wiener filter can be obtained by using a pair of equations (9) and (10) for the optimal adaptive convergence parameter and a pair of equations (9) and (11) for the heuristic adaptive convergence parameter algorithm. The results obtained are compared.

는 모든 적응성 알고리즘들에 대해 사용된다. 이러한 근사는 채널이 에지 효과들 및 멀티패스 페이딩 환경들을 겪는 도시 환경에 대해 일리가 있다. 위너 필터는 순수 AWGN 정적 환경 하에서의 최적 해이다. 그러나, 위너 해는 빈번한 에지 효과들을 겪는 채널 하에서 최선이 되지는 않는다. 또한, 퍼펙트 웨이트 벡터가 채용된다고 할지라도, 잡음 벡터의 존재 때문에 어떤 저하(degradation)가 예상된다. 본 발명의 한가지 실시예를 포함하는 시뮬레이션 테스트 환경들에 대한 파라미터들이 표1에 나열되어 있다. For comparison, other typical adaptive algorithms are also tested by simulation and compared with the present invention. For example, it is set to a least mean square (LMS), a normalized least mean square (NLMS), a recursive least square (RLS), a minimum mean square error (MMSE) proposed by the present inventors, a Choi device, and an array response vector. An adaptive algorithm with a perfect weight vector is tested through simulation. If the algorithm requires R _yy , then an instantaneous matrix

Is used for all adaptive algorithms. This approximation makes sense for an urban environment where the channel experiences edge effects and multipath fading environments. Winner filter is the optimal solution under pure AWGN static environment. However, the Wiener solution is not the best under a channel experiencing frequent edge effects. Also, even if a perfect weight vector is employed, some degradation is expected due to the presence of the noise vector. Parameters for simulation test environments that include one embodiment of the present invention are listed in Table 1.

3 and 4 illustrate the heuristic adaptive convergence parameter algorithm and the optimal adaptive algorithms in the present invention using a pair of equations {(9), (10)} and {(9), (11)}, respectively. Show each flow chart.

3 illustrates an adaptive convergence algorithm for a smart antenna processor.

In step 301, the system makes an initial guess with the weight vector w (0) set to (1, ..., 1), and fixes the Lagrange product γ to λ _max (equal to M). In step 303, a new signal is received and the post PN despread signal is observed to obtain a new signal vector y (k). Steps 305 and 307 correspond to the smart antenna processor (207 in FIG. 2). In step 305, the array output is determined by the complex array output z (k), and the optimal adaptive convergence parameter [mu] (k) is determined (according to equation (10)). In step 307, a new weight vector w (k + 1) is determined (according to equation (9)).

4 illustrates an alternative embodiment of an adaptive convergence algorithm for a smart antenna processor. This embodiment is heuristic.

In step 401, the system makes an initial guess with the weight vector and locks the Lagrange product to a constant. In step 403, a new signal is received and a post PN spread y (k) is obtained. In step 405, the array output is determined to be a complex array output z (k), and the adaptive convergence parameter [mu] (k) is determined (according to equation (11)). In step 407, a new weight vector w (k + 1) is determined (according to equation (9)).

5 shows a single realizaton of DOA estimates for the iterative (ie snapshot) index k for different adaptive algorithms. Results for LMS, NLMS, and RLS were also tested through simulations, but are not shown in FIG. These algorithms track the total angle of input including fading phase, thermal noise phase, and DOA. On the other hand, the two adaptive convergence parameter algorithms (optimum mu and heuristic mu) of the present invention, the algorithms of MMSE, Winner, and Choi in the co-pending US patent publication by the inventors can track DOAs separately. . Since the snapshot rate is set to the code symbol rate, each snapshot takes G chips to update the weight vector. The edge effect, or DOA of the desired signal, leaps from 0 ° to 40 ° in the 10th iteration. Initial weight vectors for all algorithms are set such that the estimated DOA equals 0 ° before the first edge effect occurs in the tenth iteration. 5 indicates that the smart antenna algorithm has a fast convergence rate and a small MSE.

In addition, Figure 5 shows that the two adaptive convergence algorithms of the present invention take only four iterations, while Choi's invention takes 22 iterations to reach 90% of the target DOA, such as 40 °. Point. The Winner algorithm can converge faster than the present invention, but shows large ripples in steady state. Thus, adaptive convergence algorithms such as in the present invention can be made to work better in terms of convergence rate and MSE in steady state.

Figure 6 shows two curves (optimum mu and heuristic mu) from Figure 5 for the two adaptive convergence algorithms of the present invention. Each tracking behavior of the optimal adaptive convergence parameter algorithms of equations (9) and (10) is slightly faster than each tracking behavior of the heuristic adaptive convergence parameter algorithm of equations (9) and (11). However, since the optimal μ (k) minimizes the cost function J ( w (k)) of equation (8) but does not need to minimize MSE, the MSE of the optimal convergence parameter algorithm is slightly less than the MSE of the heuristic adaptive convergence parameter algorithm. Can be larger.

7 shows the uncoded BER performance of a CDMA system with a smart antenna for the number of edge effect occurrences of a frame interval of 20 ms by using different adaptive algorithms. All algorithms have similar BER performance. The adaptive convergence algorithms of the present invention and Choi's algorithm are almost similar and slightly better than others. The BER of the algorithm with the perfect weight vector is consistent with the theoretical BER.

Table 2 is a list of the number of calculations per snapshot for each smart antenna, and M is the number of antenna array elements constructed according to one embodiment of the present invention showing a smaller number of calculations per snapshot than Choi's device. to be. Table 2 summarizes the number of complex calculations per snapshot. RLS requires the largest number of calculations per snapshot, and Winner requires the smallest number of calculations per snapshot. The number of calculations per snapshot for the adaptive convergence algorithm of the present invention is less than Choi's device.

는 본 발명의 적응성 알 고리즘들에 의해 얻어질 수 있다. 스마트 안테나가 희망 신호로부터 DOA를 추적하고 웨이트 벡터가 웨이트 벡터의 제1 요소에 의해 매 반복마다 정규화되는 경우에 웨이트 벡터 w(k)는 방정식(3)의 배열 응답 벡터 a(θ)에 접근할 것이다. 상기 DOA의 추정은 다음과 같이 얻어질 수 있다. Finally, the estimation of the DOA from the desired user signal

Can be obtained by the adaptive algorithms of the present invention. If the smart antenna tracks the DOA from the desired signal and the weight vector is normalized every iteration by the first element of the weight vector, the weight vector w (k) will approach the array response vector a (θ) of equation (3). will be. The estimation of the DOA can be obtained as follows.

(27)

(27)

는 다른 경로 빔 형성, 즉, 포워드 링크 빔 형성에 대해 채용될 수 있다. 포워드 링크는 기지국으로부터 이동국으로의 채널을 암시하고, 리버스 링크는 이동국으로부터 기지국으로의 채널이다. 리버스 링크를 통해 본 발명으로 얻어진 웨이트 벡터 w(k)는, 포워드 및 리버스 링크 사이의 캐리어 주파수 차에 따른 위상 회전을 보상한 후 기지국의 전송기로부터 희망 이동 수신기의 수신기로의 포워드 링크 빔 형성에 대해 사용될 수 있다. 그래서, 상기 포워드 링크의 성능도 본 발명에 의해 개선될 수 있다.sin ⁻¹ () is an arcsine function, ∠z is the angle of z, w ₂ (k) is the second element of the weight vector w (k) at the iteration k, and π is the radian of 180 ° Angle. DOA estimation obtained according to the present invention

May be employed for other path beamforming, ie, forward link beamforming. The forward link implies a channel from the base station to the mobile station, and the reverse link is a channel from the mobile station to the base station. The weight vector w (k) obtained with the present invention on the reverse link is used for forward link beam formation from the transmitter of the base station to the receiver of the desired mobile receiver after compensating the phase rotation according to the carrier frequency difference between the forward and reverse links. Can be used. Thus, the performance of the forward link can also be improved by the present invention.

및 실제 DOA θ사이의 MSE를 열거한다. 즉, (1) 순간적 구간들을 포함하는 시뮬레이션 테스트 인터벌들, 및 (2) 정상 상태 부분들만의 2가지 경우들이다. 정상 상태 MSE는, 추적 각이 타깃 각도의 90%에 처음으로 도달하는 반복점으로부터 다음 에지 효과까지 매 에지 효과 인터벌마다 에러들의 제곱들을 평균함으로써 얻어진다. 10, 100, 및 1000 개의 에지 효과들이 시뮬레이션 테스트들에 대해 매 20ms 마다 랜덤하게 발생된다. 본 발명 및 Choi의 디바이스는 다른 것들보다 더 작은 정상 상태 MSE를 갖고, 위너는 가장 큰 정상 상태 MSE를 갖는다. Table 3 is a list of MSE and steady state MSE for each smart antenna under edge effects including transient periods in accordance with one embodiment of the present invention, and shows comparable MSE for competing inventions. Table 3 shows DOA estimates for several different DOA tracking algorithms for the following two cases:

And MSE between actual DOA θ. That is, (1) simulation test intervals including instantaneous intervals, and (2) two cases of steady state portions only. Steady state MSE is obtained by averaging the squares of the errors every edge effect interval from the repeat point where the tracking angle first reaches 90% of the target angle to the next edge effect. 10, 100, and 1000 edge effects are randomly generated every 20 ms for simulation tests. The present invention and Choi's device have a smaller steady state MSE than others, and Winner has the largest steady state MSE.

In conclusion, a smart antenna with the optimal or heuristic adaptive convergence parameter algorithms of the present invention may be more efficient than conventional systems, especially under an edge effect environment for CDMA reverse link. The adaptive convergence parameter algorithms of the present invention show much better DOA tracking capability and slightly better bit error rate performance than conventional devices. In addition, the number of computational burdens per snapshot for the heuristic adaptive convergence parameter algorithm is less than the number of computational burdens in conventional devices.

Although preferred and optimal methods for practicing the invention have been described, those skilled in the relevant art will recognize that various alternative designs and embodiments of practicing the invention are possible and they are the scope of the claims that follow. Will fall within the scope of.

Claims (26)

A signal receiving method for using a signal receiving system in combination with wireless communication, Receiving signals with a plurality of antennas; And Processing the received signal using an adaptive convergence parameter that is adaptively changed and is employed in a smart antenna processor, The received signal is,

Where μ (k) is the adaptive convergence parameter, w ^H (k) is Hermition of the M by 1 weight vector w (k), ie conjugate and transpose, y (k) is the post PN distracted M by 1 vector, w (k) is the weight vector, z (k) is the adaptive filter output; γ is a constant value) Processed according to, And the plurality of antennas is a multiple antenna array. delete According to claim 1, And the plurality of antennas are multiple antennas. delete A signal receiving method for using a signal receiving system in combination with wireless communication, Receiving signals with a plurality of antennas; And Processing the received signal using an adaptive convergence parameter that is adaptively changed and is employed in a smart antenna processor, The signal is,

Where μ (k) is the adaptive convergence parameter, w (k) is the weight vector, z (k) is the adaptive filter output; γ is a constant value) Which is processed according to the method. According to claim 1, The processing step includes estimating a direction of arrival angle (DOA) for the antennas, wherein the DOA is separated from other signal data. According to claim 1, Determining a better weighting coefficient. The method of claim 6, Using the DOA in forward link transmission. The method of claim 6, Using DOA in reverse link transmission. According to claim 1, And the antennas are in a base station. According to claim 1, And the antennas are in a mobile station. A signal receiving system for use in combination with wireless communication, At least one signal processor having adaptive convergence parameters responsive to the signals received by the plurality of antennas and adaptively changed to be employed in the smart antenna processor, The adaptive convergence parameter in the at least one signal processor is

Where μ (k) is the adaptive convergence parameter, w ^H (k) is Hermition of the M by 1 weight vector w (k), ie conjugate and transpose, y (k) is the post PN distracted M by 1 vector, w (k) is the weight vector, z (k) is the adaptive filter output; γ is a constant value) ego, Signal receiving system further comprises a plurality of antennas. delete The method of claim 12, And the plurality of antennas comprises a multiple antenna arrangement. The method of claim 12, And a transmitter coupled to the at least one signal processor. The method of claim 12, And a receiver coupled to the at least one signal processor. The method of claim 12, And the plurality of antennas comprises multiple antennas. The method of claim 12, And the at least one signal processor comprises a filter, the filter having an adaptive convergence parameter. delete A signal receiving system for use in combination with wireless communication, At least one signal processor having adaptive convergence parameters responsive to the signals received by the plurality of antennas and adaptively changed to be employed in the smart antenna processor, The adaptive convergence parameter in the at least one signal processor is

Where μ (k) is the adaptive convergence parameter, w (k) is the weight vector, z (k) is the adaptive filter output; γ is a constant value) Signal receiving system. The method of claim 12, Wherein the at least one signal processor comprises a measurement of DOA for a plurality of antennas, the measured DOA being separate from other signal data. The method of claim 12, And said at least one signal processor comprises the determination of a better weighting coefficient. The method of claim 21, And the measured DOA is obtained from forward link transmission. The method of claim 21, The measured DOA is obtained from a reverse link transmission. The method of claim 12, And a base station, wherein the plurality of antennas are in the base station. The method of claim 12, Further comprising a mobile station, wherein the plurality of antennas are in the mobile station.

Images